Leveraging Smart Card Intelligence to Improve Transit Planning: A Case Study of PRESTO in the Greater Toronto and Hamilton Region By Tony Zhuang A thesis submitted in conformity with the requirements for the degree of Master of Applied Science Department of Civil & Mineral Engineering University of Toronto © Copyright by Yan Zhuang 2018 ii Leveraging Smart Card Intelligence to Improve Transit Planning: A Case Study of PRESTO in the Greater Toronto and Hamilton Region Tony Zhuang Master of Applied Science Department of Civil & Mineral Engineering University of Toronto 2018 Abstract While smart cards are primarily designed to automate fare collection, the transaction records contain rich intelligence. This thesis demonstrates the capabilities of smart card analytics in the Greater Toronto and Hamilton Area (GTHA), using data from the PRESTO smart card system. The case studies involve Hamilton Street Railway (HSR), GO Transit (GO), Burlington Transit, and the Toronto Transit Commission (TTC). Methods are demonstrated for: geocoding PRESTO transactions, reconstructing trips, visualizing passenger flows, spatial clustering of passenger groups, computing route level transfer volume matrices, and quantifying transfer delays. The results underscore PRESTO’s superior capability in providing longitudinal transit demand profiles, which can be leveraged to develop enhanced Key Performance Indicators (KPIs) that measure the transit travel experience from the lens of passengers. As PRESTO cards will become ubiquitous in the GTHA, this study addresses the pressing need to demonstrate the value of PRESTO analysis to improve the practice of transit planning. iii Acknowledgements First, to my co-supervisors, Professor Eric J. Miller and Professor Amer Shalaby, it has been a real pleasure being your student. My sincere gratitude goes out to both of you for the ongoing support in the last two years. I still remember that we arrived at this thesis topic of PRESTO analytics within minutes of our very first meeting back in September 2016. I was immediately intrigued and have never regretted since. Even though the last two year has not always been smooth sailing, never for a moment have I not enjoyed working on this thesis. I would like to thank my former Supervisor of Planning and Schedule at Burlington Transit, Garth Rowland, for agreeing to provide me with full access to Burlington Transit’s data that kick-started the analytics of this thesis. My gratitude also goes out to Steve Lucas, Transit Planner/Analyst at Burlington Transit, for helping me extract and interpret the data. I would also like to thank TTS 2.0, OGS, NSERC, Trapeze, OCE, and SOSCIP for funding this research, as well as Burlington Transit, Metrolinx, and Hamilton Street Railway for providing the data. To my close friends and what I like to refer to as the elite GIS strike team from the University of Waterloo, Christina Zhang, Elkan Wan, Andus Chen, and Alvin Fan, thank you for all your support and for always being a phone call (or a Facebook message) away to offer me assistance whenever I ran into difficulties. To my parents as well, thank you for all your support in the past two years. Finally, to the University of Toronto’s StarCraft II Team, it has been a real pleasure being your team lead for the past 2 years. I truly enjoyed sharing our passion for the game together and I am grateful for the support from you all to re-establish the team. During times of loneliness and anxiety, I am very grateful to be accompanied virtually via voice chats and in person. iv Table of Contents Acknowledgements .......................................................................................................................... iii Table of Contents ............................................................................................................................. iv List of Tables...................................................................................................................................... v List of Figures .................................................................................................................................. vi 1 Introduction ................................................................................................................................ 1 1.1 Background and Motivation ................................................................................................ 1 1.2 Research Questions and Objectives .................................................................................... 1 1.3 Overview of Smart Card Potentials and Past Research ....................................................... 2 2 Data Description and Case Study Overview .............................................................................. 4 2.1 PRESTO Smart Card System Overview ............................................................................. 4 2.2 Hamilton Street Railway (HSR) .......................................................................................... 6 2.3 GO Transit (GO) ................................................................................................................. 6 2.4 Burlington Transit ............................................................................................................... 7 2.5 Toronto Transit Commission (TTC) ................................................................................... 7 3 Estimating and Visualizing Passenger Flows ............................................................................ 8 3.1 Geocoding PRESTO Transactions with Historical CAD-AVL Data .................................. 8 3.2 Trip Reconstruction – Interpolating Alighting Location ................................................... 12 3.3 Development of a New Zonal System for Public Transit O-D Matrices .......................... 15 3.4 Use of Fare Zones to Aggregate O-Ds and Visualize Passenger Movements .................. 18 4 Enhancing Key Performance Indicators with User-Based Metrics.......................................... 20 4.1 Computing Transfer Volume Matrix ................................................................................. 20 4.2 Informing Service Span Changes with Assessment of User Impacts ............................... 24 4.3 Application of Spatial Clustering to Visualize Travel Patterns......................................... 28 4.4 Quantifying Travel Time Variations and Transfer Delays ................................................ 36 5 Conclusion, Reflections, and Future Opportunities with PRESTO ......................................... 44 5.1 Summary of Main Findings ............................................................................................... 44 5.2 Reflection on Software Tools ............................................................................................ 46 5.3 Reflection on Smart Card Data’s Strategic Importance .................................................... 46 5.4 Future Research Opportunities with PRESTO .................................................................. 48 6 References ................................................................................................................................ 49 v List of Tables Table 5-1: Trip reconstruction results .............................................................................................. 14 Table 6-1: Transfer volume to and from Route 1 and Route 101, Sept 2017 to Dec 2017 ............. 23 Table 6-2: Boarding and unique rider counts before and after 9:00PM .......................................... 27 Table 6-3: Profiles of riders before and after 9:00PM ..................................................................... 27 Table 6-4: Sample of classified boardings by home transit agency and TTC Stop ID .................... 29 vi List of Figures Figure 2-1: Municipalities in the Greater Toronto and Hamilton Area (GTHA) .............................. 4 Figure 3-1: Geocoding PRESTO transactions using historical AVL data ....................................... 10 Figure 3-2: Map of geocoded HSR PRESTO boarding ................................................................... 11 Figure 3-3: Near analysis concept .................................................................................................... 13 Figure 3-4: Boundaries of Traffic Analysis Zones (TAZs) along arterial roads in Hamilton ......... 15 Figure 3-5: Consolidation of timing points within 200 metres of each other .................................. 16 Figure 3-6: Transit Analysis Zones for HSR service area ............................................................... 17 Figure 3-7: GO Transit one-way O-D flows grouped by fare zones................................................ 19 Figure 4-1: Route 1 and Route 101 Express, Burlington Transit ..................................................... 21 Figure 4-2: Existing Burlington Transit service span ...................................................................... 25 Figure 4-3: Proposed coverage service span concept ...................................................................... 25 Figure 4-4: Proposed midpoint and ridership service span concept ................................................ 26 Figure 4-5: Clusters of Mi-way riders in Toronto ............................................................................ 30 Figure 4-6: Clusters of Brampton Transit riders in Toronto ............................................................ 31 Figure 4-7: Clusters of